Figure 6.5:Constant force experiments of glutaraldehyde stabilized and non-stabilized vimentin IFs presented in force-strain plots and strain-time plots in a double logarithmic scale. (A,C) About 20 individual glu-taraldehyde stabilized filaments stretched at constant forces ranging from 100 to 700 pN (see color code).
(B,D) About 100 individual untreated filaments stretched at constant forces ranging from 50 to 700 pN (see color code).
Comparing constant force experiments of untreated (Fig. 6.5 B and D) and stabilized (Fig. 6.5 A and C) vimentin filaments reveals several effects of the glutaraldehyde treatment. The most prominent one is again the "unextensibility" at small forces. While FCs at 100 pN in un-treated filaments on average yield a strain of 0.4 and the filaments are stable only for some minutes to half an hour, the tested stabilized filament yielded a final strain of less than 0.03 af-ter one hour of applied force. The effect is similar comparing the FCs at 250 pN of both filament types. After one hour of constant pulling stabilized filaments yield a strain of less than 0.1, while untreated filaments reached a strain of more than 1 in the same time. However, considering the
6.3. Creep Behavior at Constant Force 117 FCs of glutaraldehyde stabilized filaments at 500 pN, there is a clear increase in filament exten-sion which can be explained by the fact that the plateau, where theα-β-transition is presumed to happen, starts at approximately that force in stabilized filaments (Fig. 6.2 A). As the plateau in untreated filaments is reached at about 250 pN one should compare the 250 pN FC measure-ments of untreated and the 500 pN FC experimeasure-ments of stabilized vimentin. When stretched by a force that overcomes the plateau force, the final strain in stabilized filaments levels of at about 0.8 while in untreated vimentin a strain of about 1.2 to 1.4 is reached.
The step size distribution of stabilized filaments under constant load is shown in Fig. 6.6 in individual graphs per force. Again, the step sizes and the percentage steps contribute to the elongation in stabilized filaments is similar compared to untreated vimentin filaments.
Figure 6.6:Step size histograms of glutaraldehyde stabilized vimentin filaments. (A) Data from one filament at 100 pN. (B) Data from three filaments at 250 pN. (C) Data from nine filaments at 500 pN. (D) Data from five filaments at 700 pN.
6.4 Discussion
Even though the glutaraldehyde reaction is known to be reversible [1], the one and two hour force clamp measurements give evidence that this reversibility does not play a role under the given experimental conditions. Especially the 100 and 250 pN FC data do not show any effect that would indicate that glutaraldehyde is "washed out". The measurements are stable and the filaments’ response does not change over time.
An effect due to glutaraldehyde stabilization is clearly visible, especially at forces below 500 pN. However, it is not straight forward to reveal how the molecular structure is modified or rather which part of the transition or sliding is altered. Comparing all experiments it seems to be clear that the stabilization with glutaraldehyde did not freeze one of the mechanisms ex-clusively but altered the mechanical response bearing on the force level where the plateau is reached and on the maximum strain that can be reached by this setup.
One explanation for the altered mechanical response of glutaraldehyde treated filaments would be that parallel α-helices are cross-linked and require a simultaneous unfolding. This presumable leads to a higher unfolding force as observed in the experiments refering to the shifted force plateau and the reduced strain reached by FCs below 500 pN. Tight cross-linking of parts of the filaments’ monomers could explain why the final strain is reduced compared to fully unfolded untreated filaments. The comparable strain for glutaraldehyde stabilized fila-ments and untreated filafila-ments at medium to high loading rates also fits to this hypothesis as many of the cross-linking positions were found in the region of coil 1A and 1B [8] and these are also the parts that are more unlikely to unfold in untreated filaments at higher loading rates [12].
References 119
References
1. Migneault, I., Dartiguenave, C., Bertrand, M. J. & Waldron, K. C. Glutaraldehyde: Behav-ior in Aqueous Solution, Reaction with Proteins, and Application to Enzyme Crosslinking.
BioTechniques37,790–802 (2004).
2. Guzmán, C.et al. Exploring the Mechanical Properties of Single Vimentin Intermediate Filaments by Atomic Force Microscopy.J. Mol. Biol.360,623–630 (2006).
3. Herrmann, H.et al.Structure and Assembly Properties of the Intermediate Filament Pro-tein Vimentin: The Role of Its Head, Rod and Tail Domains. J. Mol. Biol. 264, 933–953 (1996).
4. Ando, S.et al.Morphological Analysis of Glutaraldehyde-Fixed Vimentin Intermediate Fil-aments and Assembly-Intermediates by Atomic Force Microscopy.Biochim. Biophys. Acta 1702,53–65 (2004).
5. Mücke, N., Kirmse, R., Wedig, T., Leterrier, J. F. & Kreplak, L. Investigation of the Morphol-ogy of Intermediate Filaments Adsorbed to Different Solid Supports.J. Struct. Biol. 150, 268–276 (2005).
6. Steinert, P. M., Marekov, L. N. & Parry, D. A. Diversity of Intermediate Filament Structure.
Evidence that the Alignment of Coiled-Coil Molecules in Vimentin is Different from that in Keratin Intermediate Filaments.J. Biol. Chem.268,24916–24925 (1993).
7. Block, J., Schroeder, V., Pawelzyk, P., Willenbacher, N. & Koster, S. Physical Properties of Cytoplasmic Intermediate Filaments.Biochim. Biophys. Acta1853,3053–3064 (2015).
8. Downing, D. T. Chemical Cross-Linking Between Lysine Groups in Vimentin Oligomers is Dependent on Local Peptide Conformations.Proteins25,215–224 (1996).
9. Fudge, D. S., Gardner, K. H., Forsyth, V. T., Riekel, C. & Gosline, J. M. The Mechanical Prop-erties of Hydrated Intermediate Filaments: Insights from Hagfish Slime Threads.Biophys J85,2015–2027 (2003).
10. Qin, Z., Kreplak, L. & Buehler, M. J. Hierarchical Structure Controls Nanomechanical Prop-erties of Vimentin Intermediate Filaments.PLOS One4,e7294 (2009).
11. Bendit, E. G. A Quantitative X-Ray Diffraction Study of the Alpha-Beta Transformation in Wool Keratin.Text. Res. J.30,547–555 (1960).
12. Block, J.et al.Nonlinear Loading-Rate-Dependent Force Response of Individual Vimentin Intermediate Filaments to Applied Strain.Phys. Rev. Lett.118,048101 (2017).
Chapter 7
Summary, Discussion and Conclusion
Single vimentin filaments show the remarkable physical properties of IFs that were already pre-sumed from mechanical measurements of IF networks and fibers. Here, based on the literature, two theoretical models, taking different levels of the architecture of mature vimentin filaments into account, were used to fit and model the collected data. The elastically coupled two-state model, adapted from references [1, 2], assumes that the parallel monomers in each ULF can be modeled as one effective element [3] yielding a system of entropic springs connected to an elas-tic module [4]. In this model it is taken into account that the vimentin monomer contains three α-helical sections that are connected by linker regions. To incorporate the observed viscous be-havior, a dashpot is introduced in the two-state model as a third element, additionally to the elastic and the entropic springs. The second, stochastic, model takes the parallel elements into account yielding a better description of the parallel elements in the ULF architecture. However in this model viscous and entropic contributions to the mechanics of vimentin are neglected and the architecture of the monomer containing three α-helical parts is not explicitly taken into account. The filament is described as a chain of ULFs and each ULF contains one spring describing the linker regions as well as the ULF to ULF connection, and 32 monomers described by their spring constant and an element that can switch between the shorter α-state and the elongated β-state. Being aware that both models are restricted concerning their application (described in Section 5.4) it was possible to relate the molecular structure of vimentin to the observed mechanical response.
The three different regimes that can be discerned in stretching curves of single vimentin IFs [4] can be attributed to an elastic response of the stretched α-helices [5–7], the transition of α-helices intoβ-sheets [5, 8, 9] and pulling on the beta-sheets [5], respectively. Even though it was not possible to directly prove theα-helix toβ-sheet transition in vimentin in the scope of this thesis, the well fitting theory (Chapters 4, and 5) as well as computer simulations [5] and
121
the direct observation of this mechanism in vimentin hydrogels [9] support this hypothesis.
Due to the loading-rate dependent behavior [4], vimentin is very soft when pulled slowly, but much stiffer in consequence of a sudden, fast load. On the molecular level this behavior can be described when the threeα-helices that form the monomers possess different probabilities to unfold and these probabilities are loading-rate dependent [4, 10].
By comparing the stretching and the retraction curve of single vimentin IFs that were stretched to about 600 pN and subsequently relaxed, a pronounced hysteresis becomes evident.
Whether the energy uptake by the filament that is equal to the area between the stretching and the relaxation curve in the force over strain graph is directly dissipated or stored in the changed conformation of the filament is not easily distinguishable. Fitting the two-state model [4] to the data indicates that the mechanism for the apparent energy dissipation is mainly the non-equilibriumα-helix toβ-sheet transition and only to a minor extend due to viscous contribu-tions (Section 5.3.1).
An additional indicator for different modes of filament elongation was found by constant force experiments. A qualitatively different elongation behavior was observed in filaments pulled at constant forces below or in the force regime of the plateau compared to filaments pulled at constant forces above the plateau force. While under low constant forces filaments mainly elongate by discrete steps, the elongation under high forces is dominated by creep (Sec-tion 5.3.3). Step detec(Sec-tion revealed step sizes that correspond well to the transforma(Sec-tion of α-helices intoβ-sheets (Section 5.3.3 and Fig. 5.3 D) whereas the creep behavior could be ex-plained by sliding of subunits against each other as suggested by simulations of the vimentin tetramer [5].
Repeated stretching-relaxation cycles, where filaments were pulled to increasing distances, revealed a complex dependency of the filaments’ mechanics on the strain history. While the filaments soften with each cycle, which is indicated by the decrease of the initial slope of each stretching event, they reach their initial length upon each relaxation (Section 5.3.2). The mech-anism for this behavior could be encoded in the parallel subunits within a filament. Assuming one ULF opens up after the other and the length of a ULF is determined only by its shortest member, elongation of the filament would be based on the fullα-helix toβ-sheet transition in ULFs. Upon force release theα-helices start to rebuild and when at least one of the parallel monomers in each ULF has changed back to itsα-helical state, the filament reaches its initial length (Section 5.3.2).
The idea to restrict parts of the mechanical properties of vimentin by glutaraldehyde stabi-lization of the filaments prior to the stretching experiment did not yield exclusive observation of one of the hypothesized mechanisms. In fact it seems to restrict parts of both mechanisms, theα-β-transition as well as sliding of subunits as the strain that can be reached by OT
mea-123 surements is decreased but the observed basic mechanisms are the same in stabilized and un-treated filaments. The differences in force and strain levels between stabilized and unun-treated filaments may be explained by cross-linking of parallel helices and the requirement of simul-taneous unfolding. Similarities between stabilized and untreated filaments stretched at higher loading rates may be explained based on literature that predicts more cross-linking in vimentin coil 1A and 1B than in coil 2 [11, 12] and the finding that loading-rate dependency may be origi-nated in the higher probability of vimentin coil 2 to unfold [4].
How exactly the physical properties of vimentin IFs found in the scope of this thesis con-tribute to the physical properties of cells, tissue and organisms is not straight forward to in-terpret. The step from single filaments to cells or even organisms is also the step from single filaments to a vimentin network that is incorporated in a complex, interactive system. Pure vimentin networksin vitroshow a similar behavior due to strain as seen in single filament ex-periments. While the network is soft at small strains, it hardens at larger strain [13]. Revealing the functions of a single protein and how it influences the others is not always straight forward and the cooperation of different proteins not necessarily exclusive. Vimentin was found to ful-fill several roles and interact with a plethora of other proteins. Disassembly of the vimentin network in 3T3 fibroblasts by microinjection of vimentin 1A peptide (a 35 amino acids long vi-mentin peptide) dramatically changes the cell shape and has a destabilizing effect on the other cytoskeletal filaments and the adhesiveness of the cells [14]. As mentioned above, Brown et al.
hypothesize that vimentin is important to control the deformability of T-lymphocyte [15]. Ad-ditionally vimentin was found to be important for the mechanical integrity of cells [16, 17] and the localization of cell components [16].
However mice with a vimentin null mutation in their germ line were found to possess a nor-mal phenotype, develop and reproduce, meaning that a complete deletion of vimentin is not generally lethal under non-pathological conditions [18]. In pathological situations it was found that wound healing in vimentin null mice is slower than in wild type mice especially concern-ing the invasion of fibroblasts and the contraction of the connective tissue [19]. The ablation of three quarters of the kidney mass was lethal by kidney failure for vimentin null but not for wild type mice [20]. In this concern vimentin might be important to modulate the vascular tone as reduction of kidney mass leads to vasodilation in the kideny in non-mutated mice. Lethality was 100 % prevented by application of bosentan an endothelin antagonist [20]. A related finding is that vimentin expression is elevated in vascular cells that are exposed to higher stress levels like the blood pressure in the pulmonary trunk or the left ventricle [21].
Based on the literature findings reviewed above and in Section 2.2 there are reasonable ar-guments to speculate about the relation between the mechanical properties of single vimentin filaments and the role of vimentin in cells and tissue. One should keep in mind that all processes
in the cell are highly complex and especially the kidney experiment [20] reveals that mechanical properties, structural changes and the biochemistry of signalling are highly interconnected. Re-garding vimentin and the vascular tone one could speculate that the structural changes in the filaments due to applied strain also enable other binding partners to bind and therefore trigger other protein pathways leading to an adaption to e.g. vasolidation or higher blood pressure of the whole cell.
Already in 1997 Galou et al. summarized that tissue lacking its normal IFs is less stable and cannot resist physical stresses as well as tissue without the IF knockout [22]. The tensile mem-ory of the vimentin filaments may lead to endothelial cells that are very well adapted, e.g., to the regular pressure shift in the vascular system due to the normal beating of the heart. The hypoth-esized structural mechanism of incompleteβ-sheet toα-helix re-transition may provide two further advantages. The first one is that filaments reach their initial length after each stretching event and therefore may prevent enlarging of the cells and the tissue. The second advantage ap-pears at larger strains, when exposed to unusual high stresses vimentin filaments, even though pre-streched for several times, may still protect the cells integrity due to the full transformation ofα-helices intoβ-sheets and subsequent stiffening.
For migrating cell types, e.g. fibroblasts, the loading-rate dependency of vimentin means that the IF network is mechanically almost invisible at low velocities like cell migration but may protect the cells when deformed very fast or to unusually large strains. In situations of extreme extension vimentin may also serve as a "shock absorber" dissipating or storing large amounts of energy byα-helix toβ-sheet transition.
To conclude, the data collected in the scope of this thesis supports the hypothesis that IFs pro-vide mechanical strength to cells and tissue and suggests an explanation why the importance of IFs becomes visible especially in pathological situations or situations of unusual mechanical load. Being aware that the models used are both limited concerning their application (described in Section 5.4), it was shown that they are consistent with each other and allow to relate the experimentally observed mechanical properties of vimentin to the underlying molecular pro-cesses. The transition ofα-helices intoβ-sheets enables vimentin to absorb or dissipate large amounts of energy and forms the basis for a mechanism of tensile memory that allows vimentin to be compliant with repeated stretching at small deformations and resistant against large de-formations. Comparing vimentin to other filamentous structures it was found to have both, the elasticity and energy absorbing/dissipating capacity of a transformableα-helix at lower strain and force values as well as the resilience and stiffness of materials like spider silk, probably due to the formation ofβ-sheets, when stretched to larger strains.
IFs are expressed in a tissue and cell-type dependent manner and are believed to support them with different mechanical properties. Based on the results collected in the scope of this
125 thesis it would now be interesting to expand the experiments on single vimentin filaments to other IF types and compare their mechanical properties to each other and also to the possible requirements of the cell types they are expressed in. Another application would be to test how the mechanical response of mutated IFs is changed compared to their wild type as mutations in IFs are known to cause several diseases.
References
1. Burte, H. & Halsey, G. A New Theory of Non-Linear Viscous Elasticity.Text. Res. J.17,465–
476 (1947).
2. Rief, M. Elastically Coupled Two-Level Systems as a Model for Biopolymer Extensibility.
Phys. Rev. Lett.81,4764–4767 (1998).
3. Friddle, R. W., Noy, A. & de Yoreo, J. J. Interpreting the Widespread Nonlinear Force Spectra of Intermolecular Bonds.P. Natl. Acad. Sci. USA109,13573–13578 (2012).
4. Block, J.et al.Nonlinear Loading-Rate-Dependent Force Response of Individual Vimentin Intermediate Filaments to Applied Strain.Phys. Rev. Lett.118,048101 (2017).
5. Qin, Z., Kreplak, L. & Buehler, M. J. Hierarchical Structure Controls Nanomechanical Prop-erties of Vimentin Intermediate Filaments.PLOS One4,e7294 (2009).
6. Qin, Z. & Buehler, M. J. Structure and Dynamics of Human Vimentin Intermediate Fila-ment Dimer and Tetramer in Explicit and Implicit Solvent Models.J. Mol. Model.17,37–48 (2011).
7. Ackbarow, T. & Buehler, M. J. Alpha-Helical Protein Domains Unify Strength and Robust-ness Through Hierarchical Nanostructures.Nanotechnology20,075103 (2009).
8. Bendit, E. G. A Quantitative X-Ray Diffraction Study of the Alpha-Beta Transformation in Wool Keratin.Text. Res. J.30,547–555 (1960).
9. Pinto, N.et al.Self-Assembly Enhances the Strength of Fibers Made from Vimentin Inter-mediate Filament Proteins.Biomacromolecules15,574–581 (2014).
10. Bell, G. Models for the Specific Adhesion of Cells to Cells.Science200,618–627 (1978).
11. Steinert, P. M., Marekov, L. N. & Parry, D. A. Diversity of Intermediate Filament Structure.
Evidence that the Alignment of Coiled-Coil Molecules in Vimentin is Different from that in Keratin Intermediate Filaments.J. Biol. Chem.268,24916–24925 (1993).
12. Downing, D. T. Chemical Cross-Linking Between Lysine Groups in Vimentin Oligomers is Dependent on Local Peptide Conformations.Proteins25,215–224 (1996).
13. Janmey, P. A., Euteneuer, U., Traub, P. & Schliwa, M. Viscoelastic Properties of Vimentin Compared with Other Filamentous Biopolymer Networks.J. Cell Biol.113,155–160 (1991).
14. Goldman, R. D., Khuon, S., Chou, Y. H., Opal, P. & Steinert, P. M. The Function of Interme-diate Filaments in Cell Shape and Cytoskeletal Integrity.J. Cell Biol.134,971–983 (1996).
References 127 15. Brown, M. J., Hallam, J. A., Colucci-Guyon, E. & Shaw, S. Rigidity of Circulating
Lympho-cytes Is Primarily Conferred by Vimentin Intermediate Filaments.J. Immunol.166,6640–
6646 (2001).
16. Guo, M.et al.The Role of Vimentin Intermediate Filaments in Cortical and Cytoplasmic Mechanics.Biophys J105,1562–1568 (2013).
17. Mendez, M. G., Restle, D. & Janmey, P. A. Vimentin Enhances Cell Elastic Behavior and Protects Against Compressive Stress.Biophys J107,314–323 (2014).
18. Colucci-Guyon, E.et al.Mice lacking Vimentin Develop and Reproduce without an
18. Colucci-Guyon, E.et al.Mice lacking Vimentin Develop and Reproduce without an